Each year the Department of Defense spends over $10 billion on its science and technology development efforts. While deemed an investment by proponents (and beneficiaries) technology development programs are particularly vulnerable in times of budget cuts. As the government moves forward with efforts to reduce spending the Department of Defense will be pressed to sustain current levels of spending on technology efforts. This situation is similar to the post-Cold War phase in defense planning when savings in spending were sought as a peace dividend. This dissertation examines the Department of Defense efforts during 1989-1992 to define certain technologies as critical to national security. Inherent in the effort to identify critical technologies was the desire to articulate technology ideology; to establish asymmetries of power and resources; and to patrol the boundaries of policy and responsibility. The questions are: What are the ideologies associated with technology development planning? What are the discursive mechanisms used to secure and reinforce power? And, what evidence of boundary work and network construction emerges from the examination? First, I distill from four years of defense technology planning documentation the explicit ideologies, the ideologies masked in metaphor, and the discourse strategies used to secure and sustain power. Following the deconstruction of the discursive elements I use Science and Technology Studies tools including boundary work, boundary objects, the Social Construction of Technology, and network theory, to further understand the heterogeneous process of defense technology development planning. The tools help explain the mechanisms by which elements of Department of Defense technology development form a connected structure. Finally, the examination yields a spherical network model for innovation that addresses the weaknesses of prior innovation network models. I conclude that in the face of uncertain budgets, technology planning relies upon ideology, power strategies, and boundary-work to build a network that protects funding and influence. In the current budget climate it will be interesting to see if the strategies are resurrected. The examination should be of interest to both the Science and Technology Studies scholar and the policy practitioner. And hopefully, the review will stimulate further examination and debate. / Ph. D.

近年來，「新公共管理」及「政府再造」已蔚為世界潮流，公共政策以民意為依歸，施政方向取決於民眾的滿意度，因此績效的衡量成為提升行政績效的基礎。我國自2002年始制定科技組織績效評鑑制度，作為各研究機構訂定評估作業之參考，對研究機構評估機制的重視可見一般。A法人為以國防科技研究發展為主的研究機構，自2014年轉型為行政法人，究竟其轉型前後其經營績效如何? 乃為本研究主要動機。 本研究以A法人所屬9個單位為績效評估的對象，運用資料包絡分析法評估各研究單位之間的相對效率，分別以專業計畫支出、服務支出、用人費用及資本支出作為投入變數；以計畫產值、營業收入、專利權等三項作為產出變數，探討行政法人化對A法人績效變動情形，以瞭解績效差異的原因，並提出改善績效之建議。 經過實證研究發現，A法人轉型後整體規模提升，究其原因主要為規模效率變動所導致；此外，相對無效率單位普遍有規模效率值較純技術效率值高的情形，顯示無效率單位應提高其純技術效率，才能提高其生產效率；再者，差額變數的分析顯示，無效率單位效率不佳的原因主要為資本支出投入資源太多，及計畫產值的產出太低，由於計畫產值涉及國防兵力整備計畫及年度預算編製等因素影響，較無法透過營運管理加以控制，因此相對無效率單位應優先考慮降低資本支出等投入資源。 / Along with the conception of "new public management" and "government reengineering" thrive around the world, every country starts to negotiate about public opinions. As important roles of satisfaction of the general public, the measurement of public policies performance has become the basis for enhancing administrative performance. We had formulated a performance appraisal system for scientific and technological organizations in 2002 as a reference for evaluating research assignments by research institutes. It means that evaluation mechanism for research institutes has become the global trend. This thesis examines the impact of administrative corporation on operating performance, not only evaluating the achievement rate of organization goal, but also promoting resource utilization efficiency. In order to evaluate the performance of the administrative corporation, this study uses data envelopment analysis to evaluate the relative efficiency among research institutes. Through the empirical result, the change of scale efficiency led to an increase in the size of the organization. Moreover, the scale efficiency is generally higher than the pure technical efficiency among relatively inefficient units. We found that inefficient units should increase pure technical efficiencies to improve their productivities. Eventually, we did confirm that there are too many resources invested in capital expenditures and output value is too low. Therefore, as output value is subject to the country budget, inefficiency units should give priority to reducing capital input to improve performance.